Options
EVE: explainable vector based embedding technique using Wikipedia
Author(s)
Date Issued
2019-08
Date Available
2019-05-15T09:26:18Z
Abstract
We present an unsupervised explainable vector embedding technique, called EVE, which is built upon the structure of Wikipedia. The proposed model defines the dimensions of a semantic vector representing a concept using human readable labels, thereby it is readily interpretable. Specifically, each vector is constructed using the Wikipedia category graph structure together with the Wikipedia article link structure. To test the effectiveness of the proposed model, we consider its usefulness in three fundamental tasks: 1) intruder detection to evaluate its ability to identify a non-coherent vector from a list of coherent vectors, 2) ability to cluster to evaluate its tendency to group related vectors together while keeping unrelated vectors in separate clusters, and 3) sorting relevant items first to evaluate its ability to rank vectors (items) relevant to the query in the top order of the result. For each task, we also propose a strategy to generate a task-specific human-interpretable explanation from the model. These demonstrate the overall effectiveness of the explainable embeddings generated by EVE. Finally, we compare EVE with the Word2Vec, FastText, and GloVe embedding techniques across the three tasks, and report improvements over the state-of-the-art.
Sponsorship
Science Foundation Ireland
Other Sponsorship
Insight Research Centre
Type of Material
Journal Article
Publisher
Springer
Journal
Journal of Intelligent Information Systems
Volume
53
Issue
1
Start Page
137
End Page
165
Copyright (Published Version)
2018 Springer
Language
English
Status of Item
Peer reviewed
ISSN
0925-9902
This item is made available under a Creative Commons License
File(s)
Owning collection
Scopus© citations
16
Acquisition Date
Mar 28, 2024
Mar 28, 2024
Views
624
Last Month
1
1
Acquisition Date
Mar 28, 2024
Mar 28, 2024
Downloads
310
Last Week
1
1
Last Month
11
11
Acquisition Date
Mar 28, 2024
Mar 28, 2024